numpy.savetxt- Save one column as int and the rest as floats?
Question:
The Problem
So I have a 2D array (151 rows, 52 columns) I’d like to save as a text file using np.savetxt. However, I want the first column’s numbers to save as integers (1950, 1951, etc) while the rest of the data saves as precision 5 (4 if rounded) floating point numbers (2.7419, 2.736, etc). I can’t figure out how to do this.
The Code
When I print the first 4 rows & 3 columns of the output of the array, it looks like this.
[[ 1950. 2.7407 2.7396]
[ 1951. 2.7419 2.736 ]
[ 1952. 2.741 2.7374]
[ 1953. 2.7417 2.7325]]
When I use the following…
np.savetxt('array.txt',data,fmt="%1.4f")
The array saves the first column as a precision 5 floating point numbers like the rest of the data (1950.0000, 1951.0000, etc). When I try to specify different formats as such…
np.savetxt('array.txt',data,fmt="%i %1.4f")
I get the following error: “ValueError: fmt has wrong number of % formats: %i %1.4f”
The Question
Is there a way I say save the first column as integers and the rest of the columns as floating point numbers?
Answers:
data has 3 columns, so you need supply 3 '%format's. For example:
np.savetxt('array.txt', data, fmt='%i %1.4f %1.4f')
should work. If you have a lot more than 3 columns, you can try something like:
np.savetxt('array.txt', data, fmt=' '.join(['%i'] + ['%1.4f']*N))
where N is the number of columns needing float formatting.
your fmt parameter needs to have the the same number of % as the columns you are trying ot format. You are trying to format 3 columns but only giving it 2 formats.
Try changing your np.savetxt(...) to
np.savetxt('array.txt',data,fmt="%i %1.4f %1.4f")
As @wflynny, but with out join:
np.savetxt('array.txt', data, fmt='%i'+' %1.4f'*N)
In my case the columns of my datamatrix were automatically concatenated, what I didnt want.
To keep the format with separate columns you can add a separator ‘;’.
The solution will look like:
np.savetxt('array.txt',data,fmt='%i' + ';%1.4f' * 4)
p.s. I was saving as .csv
The Problem
So I have a 2D array (151 rows, 52 columns) I’d like to save as a text file using np.savetxt. However, I want the first column’s numbers to save as integers (1950, 1951, etc) while the rest of the data saves as precision 5 (4 if rounded) floating point numbers (2.7419, 2.736, etc). I can’t figure out how to do this.
The Code
When I print the first 4 rows & 3 columns of the output of the array, it looks like this.
[[ 1950. 2.7407 2.7396]
[ 1951. 2.7419 2.736 ]
[ 1952. 2.741 2.7374]
[ 1953. 2.7417 2.7325]]
When I use the following…
np.savetxt('array.txt',data,fmt="%1.4f")
The array saves the first column as a precision 5 floating point numbers like the rest of the data (1950.0000, 1951.0000, etc). When I try to specify different formats as such…
np.savetxt('array.txt',data,fmt="%i %1.4f")
I get the following error: “ValueError: fmt has wrong number of % formats: %i %1.4f”
The Question
Is there a way I say save the first column as integers and the rest of the columns as floating point numbers?
data has 3 columns, so you need supply 3 '%format's. For example:
np.savetxt('array.txt', data, fmt='%i %1.4f %1.4f')
should work. If you have a lot more than 3 columns, you can try something like:
np.savetxt('array.txt', data, fmt=' '.join(['%i'] + ['%1.4f']*N))
where N is the number of columns needing float formatting.
your fmt parameter needs to have the the same number of % as the columns you are trying ot format. You are trying to format 3 columns but only giving it 2 formats.
Try changing your np.savetxt(...) to
np.savetxt('array.txt',data,fmt="%i %1.4f %1.4f")
As @wflynny, but with out join:
np.savetxt('array.txt', data, fmt='%i'+' %1.4f'*N)
In my case the columns of my datamatrix were automatically concatenated, what I didnt want.
To keep the format with separate columns you can add a separator ‘;’.
The solution will look like:
np.savetxt('array.txt',data,fmt='%i' + ';%1.4f' * 4)
p.s. I was saving as .csv